Saturday, February 28, 2015

Revisiting Ockhams Razor and broader ramifications

 A few years ago, back when I was at a loss as to why panspermia was given such a wide berth from conventional science, especially in concern with abiogenesis, my digging through Wikipedia led me to the conclusion that in choosing between Geogenesis and (other plausible theories that lead to) panspermia, parsimony was said to be invoked. Parsimony was interrelated with Ockhams razor. My perception was and still is that Geogenesis has been given no case to answer in terms of requirement for evidence. Ie. There is no evidence that abiogenesis has happened on Earth or for that matter any Earthlike planet. At the same time, the main case against panspermia is that there is "no convincing evidence" for it, and certainly no evidence for an "alternative mechanism" for abiogenesis (other than perhaps another planet with a mechanism essentially the same as what it would be on Earth) 
I thought it a bit strange that an alternative theory is sidelined completely from mainstream scientific research for the exact kind of lack of evidence that the incumbent mechanism has. No effort to Differentially tally "circumstantial" evidence between what could be a large number of plausible mechanisms and accompanying theories is ever encouraged in a scientific context. When I specifically studied the philosophy behind parsimony and Ockhams razor I came across a descriptive phrase "parsimony would usually imply a shift in the burden of proof". This described precisely what I perceived was happening in many questions between incumbent abiogenesis, evolution, astrobiology, and even astronomical object origins and any alternatives; many of which had some circumstantial evidence in their favour. 

These incumbent ideas implicitly become "new axioms" ie. They are treated as "obviously true" and quite crucially are used to come to conclusions critically dependent on these ideas. Remembering that these axioms have no evidence (any more than other ideas that are rejected) the result is that conclusions become "science" equal to that science that is constantly tested through repeatable observable phenomena.

This has been my basis of rejecting Ockhams razor, particularly for the historical sciences, which are not as subject to constant repeatable observable phenomena. 

Of course, it is not as simple as that and it has been pointed out to me that Ockhams razor is functional. Without it, one cannot reasonably perform science or engineering. Surely, if a theory is falsifiable, false theories will reveal themselves, and we can, for a time, use any plausible theory as a basis for knowledge in a field until the very moment it can be falsified. The problem to me, is that there is a great discrepancy between aspects of science that are verifiable and aspects that aren't. Aspects that are verifiable deserve no doubt to their validity. The probability that they may be completely wrong is negligible to nil. For aspects of science that are not verifiable, these are determined by a number of different philosophical techniques that may be called a great number of things - syntheses, choosing the simplest explanation that fits the data, generalising a principle from convincing anecdotes, consensus of the peer of scientists who have most deeply studied the subject, etc. the truth is, one way or another, they all amount to Ockhams razor, and they all implicitly shift the burden of proof away from the chosen theory where all other plausible theories cannot be verified  including the chosen one. The probability that the chosen theory can be completely wrong, or even impossible, is not calculable by the very nature of the phenomenon that it is explaining, but the thing that is certain is that the probability is many orders of magnitude greater than that which is verifiable.

This is the basis of my thesis that to be of better predictive value, science has to let go of Ockhams razor, and periodically shift the burden of proof back on to incumbent theories. This thesis is based on looking at science from "the outside". That is, it doesn't use one particular branch of science, to show evidence against a theory from another branch. It doesn't use observations (or amassed evidence) interpreted in the context of current incumbent theories to seed doubt on a high profile incumbent theory.

In fact, it takes an intelligent audience of outside observers who are not invested in the historical sciences, especially the status quo, to notice anything awry. Unfortunately, that limits the philosophical counter check to "Ockham razor science" to theistic philosophies with their own chosen "parsimony" which involves or invokes a God explanation explicitly or implicitly. Thus, the "negative" arguments against historical sciences, should be studied safe in the knowledge that the "positive" arguments as to  theistic explanations are doubly "Ockham razor science" and can be safely ignored.

The explanation of why it is scientifically valid, is based on "usefulness". Now if simplifications allows us predictive power, or if the maths works out better or more correctly more often that is one thing. If one is to say that heliocentricity is a "simplification" over geocentricity, this is not what I am talking about. 
If the explanation is that a simplification "works" so should be kept until proven wrong, then well it depends on the nature of the falsifiability of the synthesis in question.
Comparing different examples in history, the scientific and philosophical standard that God's omnipotence can explain one or another phenomena is at one extreme which is not generally falsifiable. Pre-tectonic plate theory geology stuck despite statistically damning evidence that continents thousands of kilometres apart were once connected. I don't see why continuing to think that places like the East Coast of South America and the West coast of Africa were "statistically independent" could possibly be "useful" in a geological sense. A theory without a verifiable mechanism that gives statistically useful results ought to be better than a (wrong) theory with an accepted mechanism that does not give those statistically useful results (also assuming that the synthesis is also not predictive or mathematical in nature), thus the former should be considered the best science can offer rather than the latter.
One question I ask of standard evolutionary synthesis is "how is it useful?", how does it "work" would a similar but importantly different statement be just as "useful" and "work" in the same way? My perception is that the mechanism's usefulness is basically self-serving to the naturist philosophy. And if you add that evolution demonstrates and clarifies the inter-relationship between all the species on Earth, mechanism independent statements are just as useful and work in the same way.


Chris Fellows said...

I find in general I am in agreement with what you say here, and see the value in having a periodic review of the historical sciences by an enthusiastic group of well-informed outsiders - who will be called all sort of names for their efforts, alas. I think valid insights from these reviews will only have a chance of 'taking', unfortunately, if they come from a place higher in the "scientific ladder" - so critiques from a solid grounding in physics might have some influence of a historical science. Do you remember Lord Kelvin's estimate of the age of the Earth?
Of course, this will leave string theory and various cosmological thought bubbles to hobble along uncorrectably for generation after generation...

I can see how a mechanism-free acceptance of continental drift would have solved problems and enable better geology, and will concede that point happily. However - at the risk of bringing things to a fevered pitch again - I don't see any gaping Atlantic Oceans in the standard model of evolution. Not to convince me, but to clarify the backstory for Marconamic Neolamarckism for future/hypothetical audiences of outside observers not personally invested in the historical sciences, would you be willing to have another go at describing why you think another model is necessary?

Marco Parigi said...

I am approaching this rejection of standard evolutionary theory from a completely philosophical stance. I want to avoid what I have done in the past - that is, get ahead of myself on what I am satisfied about the theory based on my own philosophical premises, without checking as to what people have done in the past. In this case, I had a look at what Karl Popper has had to say, and it has been very illuminating.

Marco Parigi said...

Karl Popper encourages hypotheses to be bold and falsifiable and, helpfully, encourages the *least* likely hypothesis that may fit the data. Unhelpfully, his blessing appears to be all over standard evolutionary synthesis. He could have practically authored it himself, the amount of support he has given it. Time to use his own arguments against his own pet theory.

Marco Parigi said...

Before I get to explaining why there IS gaping Atlantic Ocean wide holes in standard evolutionary synthesis, I will point out that this synthesis, if seen as a hypothesis in itself, is not bold, it is not the least likely to fit the data, and uses weasel words such as "random" to defy any realistic strategy of falsification. I am planning my full explanation in the next post rather than comments.

Marco Parigi said...

another stretch theory post

Another thing with Popper - I see little in the way of "mechanism" as requirements for new theories. In fact, there is an implied directive to think outside the box.

The evolutionary synthesis is in a reasonably strange place within the philosophy. It is considered a basis to fit other hypotheses onto rather than a hypothesis in itself. However, it is seen as bold compared to the theistic alternative predecessor. His treatise on the subject leads me to believe that the perception is of a dichotomy between creationism and evolution - one which I am claiming is a false dichotomy.
People like Wickramasinghe have actively challenged the notion that evolution as envisaged in the synthesis should just be accepted as a working model without a real avenue to falsify it.
Popper had some creative ideas, and his "spearhead" model of evolution appears to me quite removed from the consensus, and although vague in the details of the mechanism, the concept of the individual actively involved in the evolutionary process, modulated by the environment appears much closer to my initial NeoLamarckism than a pedantic reading of what the synthesis says.

Chris Fellows said...

Well, Popper's training was in cabinet-making and psychology, says wikipedia, and the rigorous fact-based nature of the first must have rebelled at the 'whatever you can get away with' nature of theorising in the latter. Neither would have encouraged him to think deeply about the importance of mechanisms.

As Feynman said, 'philosophy of science is about as useful to scientists as ornithology is to birds', and Popper is but the least worst of the learned advisors telling the authorities which swamps we live in.

Marco Parigi said...

That's it. Attack the ornithology, not the ornithologist. Yes, I realise scientists work in the real world and have to think about how they are going to get funding for research, and whether the swamp will have worms in it. I can't see how a climate scientist can get funded with a bold new theory that has nothing to do with global warming.

I really only looked at Popper because he is also looking at science from the outside, like I am. Some of the concepts I had are roughly described by him. It's not a name dropping excercise, but a way to talk about these concepts without ambiguity

Marco Parigi said...

I'm still of the thought that if a mechanism is a general thing that is required in science, and that a mountain of evidence is required to overturn an incumbent model, that this is just the sort of thing that scientific philosophers should be talking about at some point of time. It gives me an appearance that the scientists make up these rules more to protect their niche than to arrive closer to the truth. And that might is right - that a majority of scientists pooh-poohing an idea make that idea inconsequential, because they won't be funded to gather together the evidence.
With my theories that may not necessarily be a problem, as the evidence will come to us(through space research), and will mainly be in plain sight. In this case, it may not matter that nobody gives it even a fleeting chance. What is happening with comet stretch theory is a case in point. You just can't ignore evidence like that. Actually, scientists are, and it is quite disconcerting.

Chris Fellows said...

I didn't think it was a name-dropping exercise, I had just forgotten or never known what his background was!

It is axiomatic - or rather, it is a rhetorical claim made all the time in discussion of scientific revolutions - that you don't need a mountain of evidence to overturn an established theory, just one incontrovertible piece of evidence that doesn't fit. But getting everybody to notice that one piece of evidence, and getting them to agree it is as incontrovertible as you think it is... this is the thing called HARD.

Marco Parigi said...

Popper makes a criticism of "sciences" such as psychology for, essentially, being not falsifiable. Evolutionary synthesis is not falsifiable by stealth. What psychology does with theories of the mind, evolution does with rubbery definitions of every word in the synthesis. If you can fiddle with meanings of words to defend against any criticism.... I will get to a more cogent explanation in my post.

Marco Parigi said...

Strange. In my main post I didn't even get to mention anyone else's theories.

Anonymous said...

new post to hook comments on to